Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!super.upenn.edu!eecae!nancy!umix!utah-gr!spline!thomas
From: thomas%spline.uucp@utah-gr.UUCP (Spencer W. Thomas)
Newsgroups: comp.graphics
Subject: Re: Chaotic Compression
Message-ID: <2228@utah-gr.UUCP>
Date: Sun, 15-Nov-87 20:04:53 EST
Article-I.D.: utah-gr.2228
Posted: Sun Nov 15 20:04:53 1987
Date-Received: Tue, 17-Nov-87 01:16:55 EST
References: <619@applix.UUCP> <305@etn-rad.UUCP>
Sender: news@utah-gr.UUCP
Reply-To: thomas%spline.UUCP@utah-gr.UUCP (Spencer W. Thomas)
Organization: Univ of Utah CS Dept
Lines: 17

In article <305@etn-rad.UUCP> jru@etn-rad.UUCP (John Unekis) writes:
>Non information preserving algorithms wont go much over 10 to 1 on a real
>(not rigged for the algorithm) image without turning it to MUD.
> If anyone
>has seen a REAL algorithm that works better than this, please let me know.

Dr. Thomas Stockham and his colleagues, of the EE and CS departments
here at the University of Utah, have been developing vector
quantization algorithms for image compression that can (currently)
reach as little as 0.1 bit/pixel (from an original 8 bit image), a
compression ratio of 80 to 1.  The resulting picture looks a little
like what you get from a VHS video recorder on the slow speed.  It is
certainly not "MUD". 

He hopes to be able to get below 0.01 bits/pixel for motion sequences.

=Spencer   ({ihnp4,decvax}!utah-cs!thomas, thomas@cs.utah.edu)